Use this view to inspect the underlying evidence corpus. For ranked developments, decision posture and interpretation, use Signals.
Raw feed or Signals?
Raw feed is chronological evidence. Signals ranks and interprets material change.
New research proposes a unifying perspective on representation learning, framing representations as latent states capturing system dynamics under constraints.
New research proposes a dimension-free convergence theory for discrete diffusion models, addressing limitations of existing KL-based and TV bounds.
Research characterizes the VC dimension and sample complexity of Transformers, providing tight bounds for depth-L models with W parameters and sequence length T.
Research suggests long-horizon time-series forecasting benchmarks may not require complex transformer architectures, finding simpler linear models competitive.
Researchers introduced ER-JEPA, a hierarchical self-supervised learning framework for multivariate time series, demonstrated in ECG analysis.
Research explores parameter-free RDB encoders for foundation models to predict missing values across diverse enterprise tasks without retraining.
Research explores using Quantum Kitchen Sinks (QKS) for RF spectrogram anomaly detection to secure wireless spectrum management.
Research introduces Exo-MDPs, a structured class of Markov Decision Processes for sample-efficient reinforcement learning in operations research.
A survey of deepfake generation and detection techniques identifies various deepfake types, taxonomies of methods, datasets, and top detectors.
Research finds open-weight instruction-tuned LLMs consistently reproduce homogeneity bias across models, decoding settings, and identity signals.
Research proposes a method for comparing generative models using KL divergence to quantify uncertainty in model evaluation, improving rigor.
Research presents enhanced Membership Inference Attacks (MIAs) on diffusion models using a frequency-domain approach, improving data privacy breach detection.
Research explores compiling discrete programming structures, like conditionals and iteration, into differentiable recurrent neural networks.
Research proposes Minimum Bayes Risk decoding for Error Span Detection in automatic machine translation evaluation, improving error localization.
Research proposes sparse autoencoders (SAEs) to create interpretable embeddings for analyzing large text corpora, addressing limitations of LLM-based and dense embedding methods.
Research finds error amplification limits converting Artificial Neural Networks (ANNs) to Spiking Neural Networks (SNNs) for continuous control.
Researchers introduced TOPReward, a training-free method using LLM token probabilities as zero-shot rewards for general-purpose robot learning, reducing manual effort.
AG-REPA introduces a causal layer selection method for representation alignment, improving generative flow models, particularly for token-conditioned audio.
Research introduces VPWEM, a visuomotor policy using working and episodic memory to handle non-Markovian robotic tasks, addressing long-term memory limitations.
Gumbel Distillation is a new technique for parallel text generation, improving non-autoregressive model quality to rival sequential models.
Research introduces 'swarm-attack,' an open-source adversarial framework using coordinating LLM agents to bypass safety and discover software vulnerabilities.
Research introduces Pairwise Quantile Regression, extending conditional distribution modeling for high-dispersion variables beyond standard least squares.
Research introduces heat-kernel entropy profiles for weighted measures on manifolds to account for geometric support in effective sample size.
South Korean companies are investing heavily in US tech acquisitions to enhance AI capabilities and mitigate future tariff risks.
Educators report struggles with AI integration, including concerns about cognitive atrophy and faulty AI cheating detection, prompting calls for 'eval' skills.
Academic research indicates a significant increase in AI-related research volume, accompanied by a decline in overall quality, prompting concerns.
Anthropic and OpenAI are offering free and discounted tailored AI solutions to the education sector, signaling direct market engagement.
AI-powered schools claim excellence, but long-term effects on learning are unknown, and new inequalities are emerging.
Universities are developing varied strategies for AI integration, facing challenges related to costs and other policy considerations.
Universities are discontinuing AI detection tools due to persistent accuracy concerns, shifting assessment focus away from surveillance.
© 2026 OneBench: AI Insights. All rights reserved.
Evidence before opinion